J Clin Pathol: first published as 10.1136/jcp.12.1.52 on 1 January 1959. Downloaded from

J. clil. Path. (1959), 1X, 52.

THE BIOCHEMISTRY AND ANTIBIOTIC SENSITIVITY OF THE KLEBSIELLAE BY SAMUEL S. EPSTEIN* with the technical assistance of N. SPURLING Fromz thte Departmtient of Pathology of the Institiute of Laryngology and Otology, London

(RECEIVED FOR PUBLICATION DECEMBER 12, 1957)

This paper presents a biochemical survey H2S and Indole Production.-The combined SIM of a series of Klebsiellae. These findings are Difco (1953) medium was used and tested for indole interpreted in the light of current taxonomical at two and seven days. concepts. The antibiotic sensitivity of these Gelatin Liquefaction.--Liquefaction was observed organisms is also indicated. in Difco gelatin agar stabs. Nitrate Reduction. Kauffmann's (1954) medium was used. Material Cyanide Test.-The medium used was Taylor's The material analysed here comprises a total of 82 (1956) modification of that described by Vagn M0ller strains. Thirty-two come from the National (1954b). This test is based on the finding by Braun

Collection of Type Cultures and are described in copyright. Table I. The remaining 50 strains were isolated at TABLE I the Institute of Laryngology and Otology, mainly SOURCE AND SEROTYPE OF N.C.T.C. STRAINS ON from infections of the upper respiratory tract. Data KLEBSIELLA E as to their source and serotype have already been Source presented (Epstein and Friedmann, 1958). In Strain Type addition, the virulence of some of these strains for K. pneumoniae (12) I 5054 Pneumonia mice by oral and peritoneal routes of inoculation and 9494 Sputum

2 http://jcp.bmj.com/ the histopathology of pneumonic infections induced 9502 2 9503 2 Urine bv them have been studied (Epstein and Payne, 5055 Unknown 9128 8 1959; Epstein, 1959). 9129 9 9130 10 9131 11 Techniques 9132 12 9133 13 , PneumoniaNos coyput Routine Biochemical Techniques.-All tests were 3279 Untyped Nose. at 37° C. and media inoculated by Pasteur performed K. rhinoscleromat is (6) . pipettes from overnight peptone water growths. 5046 3 on September 30, 2021 by guest. Protected 5047 3 Cyanide and Koser citrate media were inoculated 5048 3 by loops and straight wires respectively. Unless 5049 3 (acapsulate form 5048) otherwise stated, all tests were read daily for seven 1936 3 days. 7799 3 carbo- Carbohydrate . Sixteen K. ozaenae (7) hydrates listed in Table 11 were tested with Andrade's 5050 4 Nose 5051 5 Unknown indicator in media consisting of lo0, carbohydrate 5053 5 (acapsulate (w/v) in peptone water. form 5051) 5052 6 Test.-Christensen's (1946) weakly buffered 8883 Untvped 'Sputum was used. 9601 Nose urea agar medium 8895 Sputum Methyl-red and Voges-Proskauer Test. The com- K. "aerogenes' (7) bined Ministry of Health (1939) oxoid media were 9527 Water used, and tests read at two days. 8844 11 Ice cream 8167 It Water Koser .-Koser's (1923) medium was 8824 II ,, 8851 II used, and positives confirmed by re-inoculation. 8801 Untyped Sputum 8808 Water *Present address: Department of Morbid Anatomy, The Hospital for Sick Children, Great Ormond Street, London. W.C.1. J Clin Pathol: first published as 10.1136/jcp.12.1.52 on 1 January 1959. Downloaded from

BIOCHEMISTRY AND ANTIBIOTIC SENSITIVITY OF KLEBSIELLAE 53

(1938) that the inhibitory effect of potassium cyanide suspended in saline, centrifuged, and the supernatants on Klebsiellae is less than on Escherichiae. again discarded. Approximately 0.5 ml. of distilled Base.-The composition is as follows: water at pH 5.0 was added, and the deposits Evans' peptone .3 g. resuspended to give suspensions of concentrations NaC .g. greater than 10' organisms/ml. (opacity tubes). As KH2PO4 .0.225 g. shown by Vagn M0ller (1954a), the permeability of Na2HP04.2H20 .5.6 g. the bacterial cell membrane to amino-acids is Aq. dist. 100 ml. increased by treatment with various lipoid solvents. The base is adjusted to pH 7.6 and then autoclaved Consequently 0.05 ml. of toluene was added to the in flasks. To 100 ml. of base, 1.5 ml. of a freshly suspensions, which were then incubated at 370 C. for prepared 0.5% potassium cyanide solution in sterile 10 minutes. distilled water is added, giving a concentration of Test. Apart from controls, seven tubes correspond- 1/13,300. The medium is distributed in 1 ml. amounts ing to the substrates were required for each strain in sterile bijou bottles and at 4' C. is stable for one under test. To each tube the following were added month. Readings are made at 24 and 48 hours. False with a 50 dropper: (1) 2 drops of suspension, (2) 2 positives may be obtained unless the caps of bottles drops of substrate, (3) 2 drops of buffer, (4) 1 drop of are screwed tightly to achieve partial anaerobiosis. indicator. The mixtures were then incubated at Gluconate Test.-The medium is a modification by 370 C. for four hours. Shaw and Clarke (1955) of that originally described Controls.-Two controls were employed with each by Haynes (1951) for the identification of Pseudo- set of tests, namely, (1) buffer, suspension, and indi- monads. The test is based on the ability of some cator (no substrate); (2) buffer, heat-killed suspension, organisms to metabolize by direct oxidation indicator, and substrate. instead of by the usual oxidative phosphorylation. Gluconate is oxidized to a reducing compound, Readings.-The pH change was read at half, one, presumed to be potassium-2-ketogluconate. The and four hours. Vagn M0ller (1955) indicated that medium is tubed in 3 ml. amounts and tested at two under these conditions of test pH change in an arginine days with Benedict's qualitative reagent. substrate can be affected by arginine dihydrolase besides decarboxylase. For this reason the specificity copyright. Malonate and Phenylpyruvic Acid Tests. Shaw of pH changes in arginine tubes was checked by and Clarke (1955) combined Leifson's (1933) sodium Nesslerization, as ammonia formation precludes malonate and Henriksen's (1950) phenylalanine significant decarboxylase activity. Random tubes in a single medium by adding phenylalan- producing an indicator change in the lysine substrate ine to Leifson's medium and supplementing with were checked chromatographically. Butanol with yeastrel. Both tests were read at 24 hours. 2N.NH4OH was the solvent and an ascending Amino - acid Decarboxylases. - The technique technique with Whatman No. 1 paper was used. described here is based on Shaw and Clarke's (1955) Spots were identified with 0.100° ninhydrin in http://jcp.bmj.com/ method. It was, however, found impossible to obtain chloroform. homogeneous suspensions from some mucoid strains Amino-acid Oxidases.-The technique is similar of Klebsiellae by washing off agar slopes, and for this to that described for decarboxylase determinations. reason suspensions were harvested instead from fluid These tests, however, were carried out at a more media. alkaline pH, and the production of a keto-acid Reagents. All seven L. amino-acids listed in Table determined colorimetricallv with ferric chloride rather III were tested at pH 5.0 and 0.03 M, except than by indicator change. on September 30, 2021 by guest. Protected glutamic acid, in which case the concentration was reduced to 0.01 M because of its strong buffer action Reagents. L. amino-acids at pH 6.8 of the same at a low pH. A 0.0125 M phthalate buffer at pH 5.0 molarity as previously were used. The buffer was was used. The indicator was 0.04% bromcresol 0.025 M phosphate at pH 6.8, and the keto-acid purple in distilled water at pH 5.0. All reagents were reagent 0.5 M ferric chloride. stored at 4° C., with a drop of toluene added to Sutspensions. - These were similarly prepared, prevent the growth of moulds. except that strains were grown in nutrient rather than Tubes.-Non-sterile 3 in. by I in. tubes were used. glucose broth and subsequently suspended in distilled Scrupulous chemical cleanliness was, however, water at pH 6.8. essential, as traces of detergent produced false positive Tests.-The mixtures were incubated at 37' C. and results. samples tested at one and 24 hours with a drop of Suspensions.-Strains under test were grown in ferric chloride. peptone water at 370 C. for four hours. Antibiotic Sensitivity.-Sensitivity to the eight anti- As maximal decarboxylase activity develops at an biotics in Table II was tested by Kohn's (1953a) simple acid pH, they were then subcultured to glucose broth diffusion technique. Sensitivity limits of organisms and incubated for 16 hours at 370 C. The broths inhibited by erythromycin, oleandomycin, and novo- were then centrifuged at 3,000 r.p.m. for five minutes biocin were estimated by the usual tube technique. and supernatants discarded. The deposits were Titrations extended over a range of 40 to 0.15 units. J Clin Pathol: first published as 10.1136/jcp.12.1.52 on 1 January 1959. Downloaded from

5454~~~~~SAMUELS. EPSTEIN

Results test, a shift of pH after half an hour is indicatix~e Table II summarizes in percentage form the of feeble enzyme activity. Feeble arginine results of the above tests. For purposes of decarboxylase activity was demonstrated in 40 convenience, the 82 analysed strains are arranged in strains. The specificity of these delayed pH the seven groups indicated. Whereas all untyped changes was confirmed by Nesslerization, with and acapsulate strains from the Institute of negative results in all cases. As is apparent, the Laryngology and Otology and the untyped majority of typed strains, including aerogenes, but N.C.T.C. strain 3279 are relegated to their with the notable exception of all but one type 3 appropriate group, the two acapsulate N.C.T.C. and more than half type 4-6 strains, possessed strains are classified with their corresponding strong lysine in decarboxylase activity. These capsular types. The captions rhinoscleroma. results are also listed in Table II for purposes of ozaenae, and aerogenes are used without prejudice correlation with the other tests. The reaction at this stage. products of randomly selected strains, exhibiting The decarboxylase activity of all strains is strong lysine decarboxylase activity, were analysed in Table III. Under the conditions of determined chromatographically and spots with Rf values equivalent to cadaverine (decarboxylated TABLE II SUMMARY OF BIOCHEMISTRY AND ANTIBIOTIC lysine) were obtained (Fig. 1). By use of the SENSITIVITY OF 82 STRAINS OF KLEBSIELLAE* combined malonate and phenylpyruvic acid medium, phenylalanine oxidase activity was detected in only one type 1 strain. Micro tests, however, proved uniformly negative. Antibiotic sensitivity results determined by a diffusion technique are listed in Table IIL Titra- No. ofstrains.. 23 5 7 9 6 25 7 tions demonstrated that all anaerogenic rhino-

Mucoid colois 91 80 85 77 100 56 0 copyright. Tenacious colonies 78 60 71 66 16 4 0 SO0LVENT Ir~N Capsules *.100 100 85 88 100 76 57 Motility .. 0 0 0 0 0 4 0 Adonitol *. 96 100 100 100 100 88 100 Inositol *. 74 80 42 100 100 52 100 Dulcitol . 13 0 14 0 16 8 28 Glucose *.100 100 100 100 100 100 100 I'lgas 91 100 14 33 83 96 85 Lactose 96 100 14 66 100 100 100

I'lgas .. 87 100 14 55 100 92 85 http://jcp.bmj.com/ Saccharose *.100 100 100 55 100 100 100 Mannitol *.100 100 100 100 100 100 100 Maltose .. 100 100 100 88 100 100 100 Rhamnose .. 100 100 85 100 100 100 100 Xylose.. .. 100 100 100 100 100 100 100 Arabinose *.100 100 100 100 100 100 101) Trehalose *.100 100 100 100 100 100 100 Sorbitol .. 100 100 100 100 100 100 100 Salicin.. .. 96 100 100 100 100 96 100 Glycerol .. 61 100 85 100 100 52 100 Starch. .. 100 100 100 100 100 100 100 .. 44 100 52 85 Urea. 87 100 14 on September 30, 2021 by guest. Protected Koser citrate .. 65 100 14 44 100 100 85 VogesProskauer 87 60 0 0 50 96 57 Methylired .. 13 40 100 100 50 4 42 Nitrate.. .. 96 100 100 88 100 100 100 Gelatin .. 0 0 0 0 0 0 0 Indole. .. 0 0 0 0 0 8 71 . .. 0 0 0 0 0 0 0 CyanideH2S .. 100 100 42 77 100 88 100 Gluconate .. 100 100 14 33 100 100 100 Malonate .. 47 100 71 11 66 100 85 Phenylpyruvic acid.. .. 4 0 0 0 0 0 0 Lysine decarb- oxylase .. 100 100 14 33 100 24 100 Percentage of strains sensitive to: Penicillin .. 0 0 0 0 0 0 0 Tetracycline .. 100 100 100 100 100 100 100 Streptomycin. 100 100 100 100 100 96 100 Chloramphenicol 100 100 100 100 100 100 100 FIG. I.-Chromatographic identification of cadaverine produced by Polysr.yxin .. 78 60 71 88 66 60 71 Erythromycin.. 0 0 85 0 0 0 0 enzymatic decarboxylation of lysine. (1) Test: Klebsiella suspen- Oleandomycin 0 0 85 0 0 0 0 sion, buffer, indicator, and lysine. (2) Control: Buffer, indicator, .. 0 0 85 0 0 0 0 and cadaverine. (3) Control: Heated Kiebsiella suspension, buffer, indicator, and lysine. (Solvent-butanol with 2N. NH40H: *Expressed as percentages of strains giving positive reactions. Whatman No. 1 paper; time of running-16 hr.) J Clin Pathol: first published as 10.1136/jcp.12.1.52 on 1 January 1959. Downloaded from

BIOCHEMISTRY AND ANTIBIOTIC SENSITIVITY OF KLEBSIELLAE 555

TABLE III AMINO-ACID DECARBOXYLASE ACTIVITY IN KLEBSIELLAE

Lysine Arginine Ornithine Histidine Tyrosine Glutamic Phenyl- No. Acid alanine

Typel .. 23 23 10 13 18 5 23 23 23 23 ,2 .. 5 5 1 4 4 1 5 5 4 1 5 3 .. 7 5 1 1 7 7 7 7 7 7 Types 4-6 .. 9 4 2 3 9 9 9 9 9 9 ,, 8-13 . . 6 6 2 4 6 6 6 6 6 Untypable and acapsulate. 25 3 9 13 7 18 10 15 25 25 10 15 20 5 "Aerogenes" .7 7 6 1 7 7 7 7 7 Total. 82 12 12 58 42 40 61 21 182 82 66 16 77 5 -t- -t:--pri unange iess inan ± nour. -t- =pt-i cnange -1-i nour- =pui cliange i-4 nours. -INSNo cfiange at '4 [iours. scieroma strains were sensitive to erythromycin, capsulated organisms, with an indole, , oleandomycin, and novobiocin over a range of Voges-Proskauer and citrate formula (1lMViC) 0.3 to 1.25 ml. with the exceptions of strain 1936 - - + +. Both intrageneric and extrageneric sensitive to 10 u. /ml. oleandomycin and S u. /ml. biochemical and serological overlap have novobiocin, and strain 7799 sensitive to 5 u. /ml. complicated the issue. Whether or not a particular oleandomycin. organism be relegated to this genus may depend on the relative emphasis placed on any particular Discussion criterion. Serologically, for instance, an organism Bergey's "Manual " (Breed, Murray, and may belong to the Escherichiae and biochemically Hitchens, 1948) divides the tribe Eschericheae to the Klebsiellae-or vice versa. In such cases into three genera, namely, Escherichia, Aerobacter, a decision as to which property represents the and Kiebsiella. Edwards (1929) challenged the aberrance may prove impossible. In an attempt to copyright. identity of Aerobacter, of which aerogenes was discourage undue reliance on any given criterion the type species, and concluded that there was no as a basis for classification, Clarke (1955) suggested justification for its extrageneric differentiation that the distribution of biochemical characters be from Klebsiellae. Kauffmann (1949), among other best described on a statistical rather than on an workers, proposed that aerogenes strains be ".6all or none " basis. Sneath (1957) has recently classified in the Kiebsiella genus. Nearly all strains lent practical emphasis to this concept by originally submitted to the N.C.T.C. as aerogenes demonstrating the application of mathematical http://jcp.bmj.com/ have recently been reclassified as K. pneumoniae computers to such taxonomical problems. (Medical Research Council, 1958). Cowan (1956) Kauffmann (1949, 1954) accepted only two in a recent survey of problems of nomenclature biochemical species, K. pneumoniae and suggested that aerogenes be accepted as a rhinoscieroma. Edwards and Fife (1955) further Kiebsiella subgroup for the convenience of water stressed that types 4-6, sometimes associated with bacteriologists, who would be loath to report on ozaena, represent a diffuse biochemical transition the of K. pneumoniae from a potable between these two species and not a discrete on September 30, 2021 by guest. Protected water. By implication, however, he accepted that entity. Henriksen (1952) proposed that ozaenae source of origin is not a valid taxonomic criterion and aerogenes species be also recognized. Other and that the designation aerogenes has thus no workers, including 0rskov (1957), recommended true status. Furthermore, Clarke and Tracey the acceptance of further biochemical subtypes. (1956) recently ascertained that high yields of The perpetuation of the name " rhinoscieroma " chitinase and cellulase, constitutive enzymes for type 3 organisms is based on tenuous evidence. supposedly characterizing many species of soil and Occasionally type 3 strains can be isolated from water including aerogenes, can also be extranasal sites. Moreover, an aetiological isolated from Klebsiellae of human origin. relationship between the organism and disease is Cowan (1954) also indicated the undesirability of not proven. It is felt consequently that such the epithet pneumoniae for the Klebsiella type strains should preferably be described numerically species. He suggested instead the name fried- by their capsular antigen. The status of motile laenderi, which does not suggest an obligate strains is still sub judice. Those liquefying gelatin relationship with respiratory infections. are best classified as Cloaca cloacae, in accordance The Kiebsiella genus was originally intended to with Kauffmann (1954), or, as Cowan (1956) include a heterogeneous group of non-motile suggests, Kiebsiella cloacae. Edwards and Fife J Clin Pathol: first published as 10.1136/jcp.12.1.52 on 1 January 1959. Downloaded from

56 SAMUEL S. EPSTEIN

(1955) recommend the designation Aerobacter Klebsiellae occasionally fail to produce gas from cloacae. Possibly non-liquefying strains represent glycerol, as did eight strains in the present series. biochemical variants, and, as Cowan (1956) No gelatin-liquefying strains were isolated, further proposes, should also be included. The although, as Lautrop (1956) pointed out, extended Institute of Laryngology and Otology strain 3502 incubation or the use of the Kohn (1953b) of the present series was such a motile non- technique may possibly unmask some liquefiers. liquefying strain. Delayed hydrolysis of urea and utilization of citrate Typical pneumoniae strains are usually aero- is characteristic of most pneumoniae strains. genic, lactose, citrate, Voges-Proskauer, malonate, Type 3 strains, with the exception of the aerogenic cyanide, gluconate, and lysine positive. Type 3 variant, were uniformly negative. Type 4-6 strains possess the reverse properties. Strains strains again behaved in an intermediate fashion, belonging to types 4-6 are sometimes aerogenic, and those utilizing citrate did so only after four citrate, cyanide, and lysine positive, and usually days. Occasional pneumoniae strains were also malonate and Voges-Proskauer negative. How- citrate and urease negative. Kauffmann (1954) ever, as indicated in Table II, biochemical pointed out that citrate-positive Escherichiae are aberrance is common in pneumoniae, especially in very rare, and that no urease-positive strains have type 1 strains. It is less frequent in type 3 strains. ever been described. Cowan (1957) has, however, One of the seven type 3 strains of the present encountered urease positive Escherichiae. Con- series was aerogenic, lactose, and lysine positive. trary to the findings of Kauffmann (1954), the In fact, although serologically this strain (No. 615) single strains of Klebsiellae types 9 and 10 studied belonged to type 3, it exhibited the biochemical in this series were citrate positive. The methyl red features of a typical pneumoniae. Henriksen and Voges-Proskauer are by no means clear-cut (1954) also described such strains and discussed tests. Results depend on temperature, duration of this problem of mutually exclusive criteria. incubation, method of testing, and, as indicated by Of the 82 strains of Klebsiellae studied in the Jennens (1954), the nature of the substrate. Of copyright. present series, 15 were anaerogenic, of which six the 89 strains of Klebsiella pneumoniae recently belonged to type 3 and five to types 4-6. As studied by Vagn M0ller (1954c), 89% were Voges- at indicated previously, gas production in type 3 Proskauer positive at one day, but only 4500 it strains can be correlated with other biochemical four days. Variables such as these make features. The majority of anaerogenic strains difficult to compare results under different all were Voges-Proskauer negative and failed to conditions of test. However, in this study type to http://jcp.bmj.com/ ferment lactose. There were no other lactose- 3-6 strains were methyl red positive. Contrary negative strains. Pickett and Cabelli (1953) Leifson (1933), who found an absolute correlation stressed the low degree of biochemical reactivity between Voges-Proskauer production and malo- of their lactose-negative strains, the majority of nate utilization, the strains studied here exhibited which, unlike those studied here, also failed to an approximate correlation only. Indole produc- The tion, while rare in Klebsiellae, is usual in ferment adonitol and inositol. strains studied by of the former and to a lesser extent of the latter are Escherichiae. Of 626 Klebsiella were fairly reliable differential tests between Klebsiellae Edwards and Fife (1955), only 26 indole on September 30, 2021 by guest. Protected and Escherichiae. Occasional strains failing to positive. " Aerogenes" type II strains are rare, ferment these carbohydrates appeared otherwise but indole positive by definition. Apart from biochemically typical. Dulcitol fermentation these, only two other strains from the Institute of and produced indole, appears rare in Klebsiellae. Kauffmann (1954) Laryngology Otology pointed out that type 10 is characteristically namely, the motile strain 3502 referred to pre- and also the acapsulate strain 3872 which dulcitol positive and this was in fact confirmed viously by the behaviour of the single type 10 strain was otherwise biochemically typical. Reduction studied. In addition, the aerogenic type 3 strain of nitrates is a non-specific test and only one nega- was also positive. In the few remaining dulcitol- tive was encountered. It is generally accepted positive strains, no other constant or significant that failure to produce H,S is a reliable exclusion biochemical deviation was observed. test for Klebsiellae ; all strains in this series were demon- (1914), on the basis of a biochemical negative. Clarke (1953) has, however, Kligler of comparison between cloacae and aerogenes strains, strated that by the use of N.C.T.C. methods maintained that gelatin liquefaction was inversely testing the reverse findings may be elicited. related to gas production from glycerol. Edwards Resistance to cyanide inhibition and the ability and Fife (1955) have shown, however, that to form reducing end-products from gluconate J Clin Pathol: first published as 10.1136/jcp.12.1.52 on 1 January 1959. Downloaded from

BIOCHEMISTRY AND ANTIBIOTIC SENSITIVITY OF KLEBSIELLAE 57 are differential tests of the utmost value. With friedlaenderi be substituted for pneumoniae as the exception of some type 3-6 strains, the great the Klebsiella type species and further that type 3 majority of Klebsiellae tested were positive. strains be no longer designated rhinoscleroma. There was, moreover, almost an absolute No evidence for any other valid species was found. correlation between these tests. It is suggested that the distribution of various The classical manometric studies of Gale (1946) properties, including biochemical, in the Klebsiella on the decarboxylases demonstrated the varied genus be best described on a statistical rather than and distinctive pattern of the enzymatic repertoire on an " all or none " basis. It is further empha- of different groups of bacteria. Subsequent sized that the results of any biochemical test must workers, notably Woiwod (1949), Vagn M0ller be qualified by reference to the particular (1954c, 1955), and Shaw and Clarke (1955), technique employed. Biochemical aberrance is introduced a variety of simplified techniques and common in pneumoniae, especially type 1 strains. stressed their practical application. Vagn M0ller K. pneumoniae is much more heterogeneous than (1954c) examined a total of 671 Enterobac- the type 3 species, of which only two subtypes, teriaceae and found that no strain, apart from the aerogenic and anaerogenic, appear to exist. K. pneumoniae, possessed powerful lysine and no Typical pneumoniae strains are usually aerogenic, other decarboxylase. The specificity of this test, lactose, citrate, Voges-Proskauer, malonate, as confirmed by the results in the present series, cyanide, gluconate, and lysine positive, and exhibit is thus high. Anaerogenic type 3 strains are delayed hydrolysis of urea. Type 3 strains, apart consistently negative and pneumoniae strains almost from the rare anaerogenic variety, possess the invariably positive. Type 4-6 strains behaved in reverse properties. Type 4-6 strains on the whole a variable manner, as did untyped and acapsulate behave in an intermediate fashion. The value of strains. Unlike Shaw and Clarke (1955), who the cyanide, gluconate, and lysine decarboxylase were able to demonstrate weak phenylalanine tests is stressed, and there is a high degree of oxidase activity in some Klebsiellae, this was correlation between them. The majority of all copyright. detected in one strain only by the combined Klebsiella strains ferment adonitol and inositol, medium and not by the micro test. but not dulcitol, fail to produce indole and H,S, All strains examined were penicillin resistant, or to liquefy gelatin. tetracycline, chloramphenicol, and streptomycin Attention is drawn to the sensitivity of anaero- sensitive, with the exception of one, resistant to genic type 3 strains to erythromycin, oleando- the last antibiotic. There is ample evidence, as mycin, and novobiocin. All other Klebsiellae are reported by Finland, Murray, Harris, Kilham, resistant to these antibiotics. http://jcp.bmj.com/ and Meads (1946) and 0rskov (1952), that resist- ance to streptomycin often develops during treat- The author would like to thank Dr. I. Friedmann ment. Approximately two-thirds of strains were for his encouragement and Dr. S. T. Cowan for advice polymyxin sensitive. There appear to be scanty in this work, Drs. J. Taylor, P. H. Clarke, P. R. data on the Edwards, and D. A. Osborn are also thanked for their sensitivity of Klebsiellae to some of help, Mr. D. Bawden for technical assistance, Mr. D. the more recently introduced antibiotics. Coppo Connolly for photographs. and Messrs. Kemball, (1957) found that the minimum inhibitory concen- Bishop & Co. Ltd., London. and Chas. Pfizer & Co., on September 30, 2021 by guest. Protected tration of novobiocin against 10 strains of " aero- New York, for their gifts of potassium gluconate. genes" ranged from 250 to 500 u./ml. Erasmus (1956) reported on seven erythromycin-resistant REFERENCES untyped Klebsiellae. Apart from the aerogenic Braun, H. (1938). Schweiz. Z. allg. Path., 1, 257. variant, all other type 3 strains exhibited suprising Breed, R. S., Murray, E. G. D., and Hitchens, A. Parker (1948). Bergey's Manual of Determinative Bacteriology, 6th ed. Williams sensitivity to erythromycin, oleandomycin, and and Wilkins, Baltimore. novobiocin. It is interesting that this previously Christensen, W. B. (1946). J. Bact., 52, 461. unrecorded sensitivity pattern is restricted to the Clarke, P. H. (1953). J. gen. Microbiol.. 8, 397. (1955). Ibid., 12, 337. anaerogenic type 3 strains, which, as suggested and Tracey, M. V. (1956). Ibid., 14, 188. previously, represent a homogeneous biochemical Coppo, A. (1957). Antibiot. and Chemother., 7, 297. subtype. Cowan, S. T. (1954). Int. Bull. bact. Nomencl., 4, 119. (1956). J. gen. Microbiol., 15, 235. (1957). Personal communication. Summary Difco Laboratories (1953). Difco Manual of Dehydrated Culture Media and Reagents for Microbiological and Clinical Laboratory A biochemical analysis of 82 Klebsiella strains Procedures, 9th ed., p. 173. Difco Laboratories, Detroit, is reported and some problems of nomenclature Michigan. Edwards, P. R. (1929). J. Bact., 17, 339. discussed. It is recommended that the name -and Fife, M. A. (1955). Ibid., 70, 382. J Clin Pathol: first published as 10.1136/jcp.12.1.52 on 1 January 1959. Downloaded from

58 SAMUEL S. EPSTEIN

Epstein, S. S. (1959). J. Path. Bact. In press. Koser, S. A. (1923). J. Bact., 8, 493. and Friedmann, I. (1958). J. clin. Path., 11, 222. Lautrop, H. (1956). Acta path. microbiol. scand., 39, 375. and Payne, P. (1959). J. Hyg. Camb. In press. Leifson, E. (1933). J. Bact., 26, 329. Erasmus, L. D. (1956). Quart. J. Med., 25, 507. Medical Research Council (1958). The National Collection of Type Finland, M., Murray, R., Harris, H. W., Kilham, L., and Meads, M. Cultures: Catalogue of Species. M.R.C. Memorandum No. 35. (1946). J. Amer. med. Ass., 132, 16. Ministry of Health (1939). Bacteriological Examination of Water Supplies (rev. ed.). Reports on Public Health and Medical Gale, E. F. (1946). Advanc. Enzymol., 6, 1. Subjects, No. 71. Haynes, W. C. (1951). J. gen. Microbiol., 5, 939. Moller, V. (1954a). Acta path. microbiol. scand., 34, 102. Henriksen, S. D. (1950). J. Bact., 60, 225. (1954b). Ibid., 34, 115. (1952). Acta path. microbiol. scand., 30, 230. (1954c). Ibid., 35, 259. - (1954). Ibid., 34, 259. -(1955). Ibid., 36, 158. Jennens, M. G. (1954). J. gen. Microbiol., 10, 121. 0rskov, I. (1952). Ibid., Suppl. 93, p. 259. Kauffmann, F. (1949). Acta path. microbiol. scand., 26, 381. (1957). Ibid., 40, 155. - (1954). , 2nd ed. Ejnar Munksgaard, Pickett, M. J., and Cabelli, V. J. (1953). J. gen. Microbiol., 9, 249. Copenhagen. Shaw, C., and Clarke, P. H. (1955). Ibid., 13, 155. Kligheer, I. J. (1914). J. infect. Dis., 15, 187. Sneath, P. H. A. (1957). Ibid., 17, 201. Kohn, J. (1953a). J. clin. Path., 6, 243. Taylor, J. (1956). Personal communication. (1953b). Ibid., 6, 249. Woiwod, A. J. (1949). J. gen. Microbiol., 3, 312. copyright. http://jcp.bmj.com/ on September 30, 2021 by guest. Protected